By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
Invest Like the BestHow He Built the World’s Largest Digital Bank with 140 Million Customers
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
CHAPTERS
- 0:00 – 1:26
Nubank at 140M customers: cost advantage before AI
David Vélez contrasts Nubank’s scale and efficiency with traditional banks, arguing that digital-first infrastructure creates a major cost-to-serve advantage even before AI. Patrick O’Shaughnessy sets up the core question: why the fintech transformation is still early despite Nubank’s growth.
- •David Vélez says Nubank grew to 140 million customers with about 10,000 employees.
- •He compares this to incumbent banks with roughly 100 million customers, far more employees, and thousands of branches.
- •He claims Nubank’s cost to serve a customer is about 3% of a traditional bank’s, and emphasizes this is “pre-AI.”
- •Patrick notes Nubank’s jump from 35 million to 140 million customers since their last conversation.
- 1:26 – 3:59
Why banking’s “Blockbuster-to-Netflix” shift is still in the early days
David argues global consumer financial services remain dominated by incumbents and branch-based distribution. He frames fintech and neobanks as still a small share of a massive industry, with a long runway for market-share migration and new customer inclusion.
- •David estimates global consumer financial services as a roughly $10 trillion industry, with about 90% still going through branches.
- •He says all neobanks and fintechs combined are still under 10% of the total market.
- •He compares banking’s current state to a world where most people still rent from Blockbuster.
- •He describes Nubank as unusually scaled in Brazil, reaching a large share of the adult population and becoming primary accounts for many customers.
- •He adds that market expansion may also come from billions of people entering the banking system over time.
- 3:59 – 5:54
Phase one: smartphones removed branches and unlocked a massive efficiency gap
David explains how smartphones enabled banks to serve customers without physical infrastructure, creating a structural cost advantage that can be passed to customers. He then positions AI as a second multiplier on top of the mobile revolution.
- •David says smartphones were the first major catalyst enabling digital banking to outperform branch-based banking.
- •He reiterates the employee-and-branch comparison to show the cost of real estate and physical infrastructure.
- •He argues this advantage can be returned to customers via lower fees, better rates, and improved experience.
- •He says AI can further multiply efficiency and expand what’s possible at scale.
- •He describes a strategic choice: use AI to run far leaner operations or raise ambition and expand broadly.
- 5:54 – 14:39
AI’s promise in banking: access, cost, and “intelligence” for everyone
David organizes banking problems into three buckets—access, cost, and intelligence—and claims AI can transform each. He argues the biggest leap may be making sophisticated financial guidance broadly available through AI rather than education programs.
- •David’s three buckets are access, cost, and intelligence.
- •On access, he notes billions lack credit access and cites low credit card penetration in some countries.
- •On cost, he highlights hidden consumer costs like low deposit yields, fees, and expensive remittances.
- •He argues AI can deliver financial “intelligence” broadly, replacing the need for traditional financial education.
- •He describes AI as a way to compete away profit pools that persist due to consumer information gaps.
- 14:39 – 17:40
Moats in the AI era: regulation, licenses, data scale, and culture
Patrick asks how a new AI-native competitor might attack Nubank, and David outlines what advantages matter in banking. He argues speed helps early startups, but regulation and data become decisive constraints, and incumbents without the right systems may still waste their scale advantage.
- •David says AI-native startups can move fast early, especially before being regulated.
- •He argues banking licenses and regulator timelines can’t be compressed simply with AI tools.
- •He emphasizes Nubank’s scale and data depth as an advantage for credit and collections decisions.
- •He says Nubank built a proprietary foundation model using its own data and describes data as a moat against new entrants.
- •He cautions that scale alone isn’t enough; systems and data culture determine whether scale becomes an advantage.
- 17:40 – 23:29
Nubank’s AI timeline: machine learning since 2013 and foundation models via Hyperplane
David traces Nubank’s use of modeling back to its founding, when it sought to price credit risk better than incumbents. He describes acquiring Hyperplane and building a proprietary foundation model for credit, later expanding its use to other decisions and markets.
- •David says Nubank has used machine learning since the company started in 2013.
- •He explains Nubank chose to take credit risk on its balance sheet and build model-driven decisioning early.
- •He describes acquiring Hyperplane, a team working on foundation models for financial services.
- •He says Nubank built a proprietary foundation model using its own data for credit decisions and published research.
- •He notes the model is also used for areas like fraud and supports launches in new markets.
- 23:29 – 32:30
The future of money rails: cash decline, open finance, and crypto/stablecoins
David lists major external shifts he’s watching: the ongoing decline of cash, the spread of open banking/open finance, and the growing role of crypto and stablecoins. He argues these changes reshape data creation, competition, and how people hold value across borders.
- •David predicts continued movement away from cash, often driven by real-time payment systems.
- •He says cash digitization increases transaction data and can expand credit access.
- •He frames open finance as consumers owning and porting their data, reducing inertia that benefits incumbents.
- •He describes banking as oligopolistic in many countries and expects more competition as data portability expands.
- •He discusses crypto and stablecoins as part of parallel systems alongside fiat, especially where trust in local currencies is low.
- 32:30 – 36:01
From Brazil to Mexico and beyond: what does and doesn’t transfer across countries
Patrick asks how portable Nubank’s learning is across markets, and David says it’s more transferable than expected. He highlights cash-heavy economies as initially harder for a digital-only player, but argues digitization makes markets more similar over time.
- •David says learnings across countries have been more exportable than he expected.
- •He describes the challenge of competing in more cash-heavy, less-banked markets without branches or ATMs.
- •He says Mexico was initially harder because more transactions were offline, but digitization narrows the gap.
- •He argues the future of financial services may look more homogeneous across countries, like other global digital products.
- •He notes Nubank has become a large digital player in Mexico while the environment becomes more digital.
- 36:01 – 41:36
Global expansion and a major announcement: US launch and stablecoin-backed accounts
David reframes Nubank’s thesis as global rather than region-specific, then explains the company’s move into the US and a stablecoin-backed account across many countries. He positions these steps as the next phase after proving the model in Latin America.
- •David says Nubank’s original thesis was that global retail banking will be transformed by digitally native companies.
- •He announces launching the US as a market and launching stablecoin-backed accounts across many countries.
- •He cites Nubank’s performance in Latin America as proof points for the model’s viability.
- •He describes the US as competitive but still dominated by incumbents and branches.
- •He identifies building a strong brand and trust in the US as a key challenge.
- 41:36 – 48:08
Learning from Amazon and Disney: building customer obsession and “magic” in banking
David describes how Amazon and Disney influenced Nubank’s approach to customer experience. He argues true customer obsession requires acting on the value even when it’s financially inconvenient, and he connects this mindset to how companies might misapply AI by focusing only on efficiency.
- •David says Nubank has been inspired by Amazon’s customer obsession since early on.
- •He explains Disney’s influence as designing “magical” experiences, even for a mundane product like banking.
- •He gives examples like the experience of receiving a card or welcome kit as a moment to create delight.
- •He argues the tradeoff of customer obsession is sticking to it when short-term profit temptations arise.
- •He notes AI conversations often emphasize efficiency more than customer experience, which he sees as a gap.
- 48:08 – 59:47
Reinventing Nubank with AI: workflow redesign, flatter org structure, and new management systems
David describes a turning point where newer models began working well for their stack, prompting an organization-wide push to rebuild workflows with AI. He explains how this effort contributed to reducing layers, increasing CEO span of control, and improving access to “ground truth” through better information systems.
- •David says Nubank shifted from remote to hybrid and sees in-person collaboration as more important, even with AI.
- •He recounts a moment when updated models began performing well for their programming needs, triggering a “code red.”
- •He says Nubank mapped and rebuilt many workflows across the organization using AI tools and evaluation infrastructure.
- •He describes flattening the org structure, reducing layers, and increasing his number of direct reports.
- •He explains staying close to ground truth through fewer layers, better internal information systems, and a culture of directly engaging the people doing the work.
- 59:47 – 1:13:18
Founder mindset and leadership: motivation, imposter syndrome, and balancing founder vs CEO modes
David reflects on what keeps him motivated, describing impact and the drive to push through major shifts like AI. He discusses imposter syndrome, the founder-versus-CEO tension, and what he believes makes great leadership—vision, persuasion, and humility to adjust.
- •David says his motivation centers on impact and that scaling Nubank further can be his highest-impact path.
- •He describes himself as more energized during disruption, likening it to a “wartime general” mode.
- •He discusses imposter syndrome as a recurring challenge as the organization grows in complexity.
- •He contrasts founder traits (impatient, fast-moving) with CEO traits (systematic, patient, building systems).
- •He defines great leadership as having a differentiated vision, inspiring others, and balancing conviction with listening and humility.
- 1:13:18 – 1:20:34
Nubank’s five-year bear vs bull cases, and a closing lesson on entrepreneurship and mission
David outlines a bear case centered on saturation in current markets and failed internationalization, and a bull case where Nubank rides technology tailwinds into many countries. He closes with an entrepreneurship lesson about overcoming imposter syndrome through deep personal commitment to a mission that sustains long-term effort.
- •David’s bear case: saturation in current countries, slowing growth, and unsuccessful international expansion.
- •He links risk to failing to capitalize on AI and other tailwinds that could enable scaling.
- •David’s bull case: significant share shift from incumbents to digital-native firms accelerated by payments rails, open finance, and AI.
- •He reiterates Nubank’s cost advantage as a key lever to return value to customers via yields, fees, and experience.
- •He says founders should seek a mission they deeply care about, because scaling is painful and requires conviction over time.
